Nicolas Silva
Nicolas Silva
@cwfitzgerald how does the queue fit in the above diagram? The WebGPU spec says that the device has a strong reference to the queue (but we can make it work...
From a glance at the code it looks like `enumset` uses `u128` if you need more than 64 bits, and then arrays of `u64` if you need more than 128...
Due to the issue with `u128` and ffi my suggestion is to split into the standard flags and native-only ones into two types. We are pretty safe with 64 bits...
I just tried to see if I could alter `enumset` to produce something that would work over ffi ( I don't think getting it to use arrays of u64 would...
Note that pipeline caches have the potential of having an even greater impact for those that are using the d3d12 backend and are stuck with FXC for one reason or...
Another instance of the same crash in `cts/webgpu/api/validation/encoding/cmds/copyTextureToTexture/cts.https.html?q=webgpu:api,validation,encoding,cmds,copyTextureToTexture:copy_ranges_with_compressed_texture_formats:*`. This time while creating a texture (but most of the stack looks the same): ``` 0 D3D12SDKLayers.dll + 0x559a6 rax = 0x0000000000000001...
This PR contains https://github.com/gfx-rs/wgpu/pull/5242
This would also fix our current issue with leaked deduplicated bindgroup registry slots.
Before we add back a complicated and/or invasive system, I'd like us to map out and document what the locking story is like in `wgpu-core`, see how many long lived...
@cwfitzgerald looks like it might have been caused by the bindgroup layout dedup refactor ?